Abstract Previous studies on the use of temperature function integration as a means of predicting the shelf-life of chilled flesh foods are reviewed and the possible application of this principle to poultry-meat products is discussed. Consideration of those mathematical models which are currently available for predictive purposes suggests that the square root equation of Ratkowsky et al. (1982) is the most appropriate according to tests on 28 sets of spoilage data from 14 published studies on poultry meat covering a temperature range of −2 to +25°C. However, discrepancies remain, particularly at higher ambient temperatures, which are relevant to abuse conditions. Deviations of observed data from the predictive models may be expected because of the apparent assumption that meat spoilage over a wide temperature range is principally a reflection of the temperature effect on growth of a single type of psychrotrophic bacterium ( Pseudomonas). In fact, spoilage at higher temperatures would appear to involve other organisms with different growth characteristics as well as a separate effect of temperature on the metabolic activities of the organisms concerned and hence on the production of spoilage metabolites. These possibilities are discussed in the light of available data. Factors apart from temperature which are known to affect the shelf-life of raw poultry products stored aerobically at high relative humidity include the type of muscle present and the initial level of product contamination with spoilage bacteria. The successful application of devices for temperature function integration in the poultry industry would appear to require both carefully standardized tests for product acceptability and rigorous control of spoilage bacteria in the processing plant to reduce existing variation in carcass contamination.